Diagnosis of maxillary sinusitis in Finnish primary care. Use of imaging techniques

Transcription

1 of maxillary sinusitis in Finnish primary care. Use of imaging techniques Marjukka Mikela' and Kirsiliina Leinonen2 ' National research and development centre for welfare and health, Helsinki, * Helsinki Health Centre, Finland Makela M, Leinonen K. of maxillary sinusitis in Finnish primary care. Use of imaging techniques. Scand J Prim Health Care 1996;14: Objective - To observe the effect of imaging techniques on the diagnostic pattern of sinusitis in primary care. Design - A multicentre survey. Setting - 14 health centres with varying facilities for clinical imaging. Subjects adult patients with a suspicion of acute maxillary sinusitis. Ultrasound or radiography facilities existed in 337 cases while 109 could only be judged clinically. Main outcome measures - The use of ultrasound, radiography, laboratory tests, irrigation and control visits. The final number of patients with sinusitis in different facility groups. Results - When available, ultrasound was used in % and radiography in 6-32 % of cases. The ultrasound finding showed mucosal thickening or was difficult to interpret in every third case; one half of these were interpreted as sinusitis and the rest were considered healthy. In total, sinusitis was diagnosed in 8448% when ultrasound or radiography only could be used, and in 77% when both techniques were available (p<o.ol). Simultaneously the general practitioners' confidence in the correctness of their diagnosis increased from 39% to 66%. Conclusion - When possible, ultrasound is widely used in diagnosing sinusitis in Finnish primary care. The use of ultrasound slightly diminished the numbers of sinusitis diagnoses but the techniques of using and interpreting ultrasound findings need to be improved. Key words: sinusitis, ultrasound, primary care. Kirsiliina Leinonen, MD, Steniuksentie 6 C 30, FIN Helsinki 32, Finland. The diagnosis of maxillary sinusitis in primary care can be based on clinical examination only or supported by imaging techniques, such as x-ray and ultrasonography. Some otorhinolaryngologists also consider imgation a useful diagnostic method (1-3). The primary care setting often lacks facilities for a precise diagnosis, and general practitioners (GPs) have had to rely on their clinical judgement despite the uncertainty involved (1,4). Radiography is usually considered reliable for showing secretion in nasal cavities (5,6), while the opinions on the accuracy of the ultrasound vary (3,5,7). Ultrasonography has been promoted as an easy and readily accessible method in primary care and it is used increasingly. Most studies have concentrated on the accuracy of radiography or ultrasonography. A few studies ( 1,4,8) have looked for a relationship between Scand J Prim Health Care 1996: 14

2 30 M. Makela. K. Leinonen clinical signs and sinusitis. Linle or no information is given of the actual utilization of the methods. Two studies in Swedish primary care (9,lO) describe the impact of introducing ultrasonography in general practice in limited geographical areas. One Norwegian study (11) has reported GPs opinions of importance of different factors in diagnosing sinusitis. In Finland, GPs often have access to either radiography, ultrasonography, or both. We report our findings of how the different technical equipment provided by health centres affected the primary diagnostic patterns in maxillary sinusitis. Material and methods Fourteen health centres participated in the study, representing the different geographical areas of the country. Both urban and rural practices were included. The number of physicians who agreed to contribute patients was 161, but individual physicians referring patients were not identified. The availability of diagnostic facilities varied between health centres. Ultrasonography (Sinuscan 101, Ultramax) was available in the health centres at all times; x-rays could usually be taken during office hours. Three health centres could use only x-rays, two only ultrasound, seven had both facilities, and two had neither. Maxillary puncture and irrigation were technically possible in every health centre at all times. The study period lasted for six weeks during autumn Patients were eligible for the study if either the patient or the physician suspected sinusitis. Adults (16 years and older) consulting for symptoms of less than 30 days duration were included. Pregnant women and patients with chronic sinus problems were excluded. During the consultation physicians filled in the questionnaire covering anamnestic data, symptoms and their duration, signs on physical examination, diagnostic methods used, the diagnosis and the certainty of this diagnosis, treatment, control visits ordered, sick leaves, and referrals to specialists. The accuracy of the data was checked against patient records when necessary, and analysed using the t- test, chi-square test, and chi-square test for trend. Three different sets of criteria for sinusitis (ultrasound, irrigation and clinical impression) were used in this study. The ultrasound and irriga- Scund J Prim Health Cure 1996; I4 tion results were evaluated separately for both sinuses. Ultrasound findings were classified in four categories: sinusitis, mucosal thickening, healthy, and difficult to interpret. The availability of imaging techniques was analysed as an independent variable in the diagnostic process, but the behaviour of individual physicians was not evaluated. The irrigation result was either positive (mucous or purulent fluid from sinuses) or negative. The <<final diagnosis, was based on all available information about the patient and was divided in three categories: certain sinusitis, possible sinusitis, and no sinusitis. Results The physicians returned 502 questionnaires. 56 were excluded because the symptoms had lasted over 30 days, the duration of symptoms was unknown, or the patient s age was less than 16 years or unknown. 446 episodes met the inclusion criteria. Imaging technology was available in 337 cases (x-ray in 60, ultrasound in 101, and both in 176 cases). 109 patients could be judged only clinically. The patients in each diagnostic facility group had a similar age-sex distribution. The background data, symptoms, and signs are shown in Table I. Pain and nasal obstruction were the most frequent symptoms, each affecting over 314 of the patients. Patients in the no technology group differed slightly from others. They suspected sinusitis themselves more often (89%), their mean duration of symptoms was shorter (6.9f6.1 days), they experienced more pain radiating in teeth (31%), pain on bending (62%), and had more unilateral signs (34%) on physical examination than patients in other groups. Diagnostic process The diagnostic process in the groups with different diagnostic facilities varied only in the utilization of the imaging techniques and the performance of irrigation. X-rays were taken in 32% of the cases when only radiography was possible, while 92% of the patients were examined with ultrasound in settings where only ultrasound could be used. When both facilities existed, x-ray was used for four (2%), ultrasound for 139 (79%), and both for six (3%) patients. The irrigation rate

3 Table I. Background data, symptoms and signs of the patients with suspected sinusitis. Anamnestic data Women Mean age, years Patient suspecting sinusitis Previous sinusitis Previous irrigation Allergic rhinitis Antibacterials < 1 month Mean duration of symptoms, days Symptoms Pain in cheeks or forehead Nasal obstruction Cough Purulent rhinorrhoea Other symptoms Unilateral symptoms Pain radiating in teeth Hyposmia Bad smell in the nose Fever over 38 Celsius All N46 (%) 328 (74) (77) 337 (76) 134 (30) 76 (17) 55 (12) 9.lf (79) 339 (76) 2 15 (48) 2 13 (48) 104 (23) 122 (27) 98 (22) 76 (17) 57 (13) 26 (6) Signs Mucosal thickening 269 (60) Pain on palpation 232 (52) Pain on bending 213 (48) Unilateral signs 113 (25) Purulent secretions in the nose 110 (2-5) Purulent secretions in pharynx 61 (14) Purulent cough 26 (6) was highest when both ultrasound and x-ray were available. 17 of 176 patients (10%) were irrigated when both imaging facilities existed and only one (in the x-ray group) or two (in the ultrasound and no technology groups) in other settings (p<o.oi). Control visits within three days were scheduled for &8 patients per group, and no statistically significant differences could be shown. Laboratory tests were rare, from one to three tests per each group. Diagnoses sinusitis in primary care 3 1 sinusitis more commonly than those examined with technology (87% compared to 73%, x2=5.6, p<0.05). Likewise they more often experienced pain on palpation (62% compared with 47%, x2=4.8, p<0.05). 70 out of the 76 patients (92%) were diagnosed as having maxillary sinusitis, compared with 263 of 337 (78%) for whom technology had been used (x2=6.6, p<o.oi). Radiography alone or together with ultrasound was available in 236 cases, but only 29 patients were x-rayed. Ten patients in the x-rayed group had received antibiotic treatment during the previous month, compared with 26 patients in the not x-rayed group (x2= 7.8, p<o.ol). Although the mean duration of symptoms in x-rayed patients (1 1.9i7.0 days) did not differ statistically from others (937.1 days), both the median and the mode were much higher in the former group (both 14 days) than in the latter one (both 7 days). Outcomes The availability of imaging facilities was associated with smaller numbers of sinusitis diagnoses and with the GPs greater confidence in the correctness of their diagnosis (Table II). Use of antibacterials diminished simultaneously with sinusitis diagnoses. Irrigated patients 22 patients were treated with sinus irrigation. Half of them had previously been treated with antibacterials, compared with 10% of those not irrigated (x2=26.8, p<o.ool). Their symptoms had lasted slightly but not significantly longer (12.4k10.3 days vs. 9.0k6.9). Two of the imgated patients were from the group in which no diagnostic technology was available. 15 patients had been examined with ultrasound, two with x-ray, and in three cases technology was available but not used. In nine patients, the imgation result was negative. Irrigations accumulated in the group where both radiography and ultrasound were available. A positive finding on ultrasonography did not necessarily give a positive irrigation result (Table 111). Diagnostic process and clinical features X-ray or ultrasound facilities existed in 337 cases. Available technology was not utilized for 76 patients. These patients had a previous history of Control visits 50 patients were asked for a control visit. In 14 of the 19 cases seen within three days, radiography was planned at control. X-rays were planned for Scand J Prim Heufrh Care 1996; I4

4 32 M. Makela, K. kinonen Table N. Outcomes in groups with different diagnostic facilities. Maxillary sinusitis - of these diagnosis certain (%) Treatment Antibacterials Control visits, all Sick leave **P<O.Ol No technology X-ray Ultrasound Both N=109 (%) N=60 (%) N=101 (a) N =176 (%) P 98 (90) 53 (88) 85 (84) 136 (77) ** 38 (39) 23 (45) 49 (58) 90 (66) ** 97 (89) 54 (90) 84 (84) 136 (77) ** 13 (12) 3 (5) 16 (16) 18 (10) 32 (29) 22 (37) 19 (19) 4 (25) Table ZII. Irrigation results and interpretation of ultrasonography (N=27 sinuses). rnucosal difficult to sinusitis thickening interpret healthy all Inigation positive Inigation negative All sensitivity 67% - specificity 92% 25 patients at control examinations. Irrigation at control was planned mostly for those who were irrigated already at the first consultation (7 of 11 patients). Interpretation of ultrasound 238 patients (476 sinuses) were examined with ultrasonography. The sound waves produce different kinds of echograms and patterns of signlights when passing through the sinuses, depending on their contents. Fluid in sinuses gives a backwall echo, whereas in airfilled sinuses the sound waves pass only to the anterior wall of the sinus. An echo between these two indicates mucosal thickening, a cyst, or some other pathology. That is why the results were classified into four groups: sinusitis, mucosal thickening, healthy, and those difficult to interpret. Table IV shows how the physician s interpretation of the ultrasound finding was associated with the final diagnosis. The interpretations and the final diagnoses are given for each examined sinus. In 306 sinuses (64%), the clinical diagnosis and ultrasound fmding coincided. Table V illustrates how different cut-off points change the sensitivity and specificity of ultrasound results with respect to the physicians final diagnosis. When all final sinusitis diagnoses are clustered, the ultrasound results can first be grouped to sinusitis or no sinusitis (Table Va). Here the specificity of ultrasound would be high, but many false negatives would mean low sensitivity. Next, when healthy ultrasound findings are considered against all others (Table Vb), the sensitivity rises to 96% as specificity drops. The best result would be achieved by accepting all clear ultrasound findings as indicative of certain sinusitis and others as negative (Table Vc). Then the diagnostic accuracy would be as high as 94%. Discussion The age- and sex-distribution of our sinusitis patients were similar to those in other studies (1,4,8). The commonest symptoms and signs (facial pain, nasal obstruction) were similar to previous findings, but these are correlated with sinusitis only marginally, or not at all (1,4). Symptoms Scad J Prim Health Care 1996; 14

5 ~~ Table N. Interpretation of ultrasound (N=476 sinuses1238 patients). Diagnoses sinusitis in primary care 33 mucosal difficult to Final diagnosis of the patient sinusitis thickening interpret healthy all Sinusitis, certain (%) 186 (90) 11 (11) Sinusitis, possible (%) 17 (8) 35 (34) 20 (63) 12 (9) 84 No sinusitis (%) 5 (2) 58 (55) 12 (37) 120 (91) 195 All (%) 208 (100) 104 (100) 32 (100) 132 (100) 476 Table V. Sensitivity and specificity of ultrasound findindgs with different cut-off points. Va) Sinusitis No sinusitis Sinusitis No sinusitis Sinusitis Other Sensitivity 203/281 = 72% Specificity 190/195 = 97% not healthy healthy I Sensitivity 269/281 = 96% Specificity 120/195 = 62% 344 I Sinusitis Other Sinusitis, certain Sensitivity 186/197 = 94% Other Specificity 251f279 = 92% or signs strongly related to sinusitis, such as unilateral facial discomfort, purulent rhinorrhoea, hyposmia, or fever, were rarely seen in secondary care studies (1,4), and even more infrequently in our primary care material. Patients in a primary care study from the Netherlands (8) had a very similar distribution of symptoms and signs to our patients. We did not gather data on patients who refused to participate in the study or who where excluded for other reasons, such as diagnostic uncertainty or pressure of time. In addition, patients in the no technology group differed slightly from the others, but in general the differences were small, sometimes contradictory (in predicting true sinusitis), and not large enough to have major clinical significance (1,4,8). For both reasons, the application of our findings to all primary care sinusitis patients must be done with caution. Moreover, Finnish physicians tend to regard sinusitis as a bacterial disease, instead of differentiating between serous and purulent forms. This Scund J Prim HeuM Cure 1996; 14

6 34 M. Makela, K. Leinonen differs from the practice in many other countries (3,8,11). Comparisons with primary care sinusitis patients in other countries are difficult also because the patterns of self-care, consultation, and referral vary from country to country (1 2). Ultrasound was used widely and significantly more often than radiography, but not universally. Most health centres had only one and at the most four ultrasound devices in their use, so not everybody could have the equipment in his or her office. The convenience of immediate availability of ultrasound may be important for its utilization. Although ultrasonography seems to be a well accepted tool in general practice, there are problems in interpretation. In 136 of the studied sinuses (29%) the findings were inconclusive. These were interpreted sometimes as diseased, sometimes as healthy. The low sensitivity of ultrasound in imgated patients may also reflect the difficulties in interpretation. Finally, some patients with negative ultrasound findings were medicated, and others with apparent sinusitis did not receive antibacterials. Similar irrationality in the use of diagnostic tests has previously been shown in the treatment of tonsillitis in general practice: taking the test had little effect on physicians actions (13). In general, the utilization rate of radiography was low. If radiography was the only method for clinical imaging, an additional reason for radiography concerning a simple suspicion of sinusitis seemed to be necessary. Most often this was preceding antibacterial treatment and longer duration of symptoms. In contrast to the primary consultation, radiography was the most popular examination at control visits. Unexpectedly, ultrasound availability did not diminish the need for control visits within three days, and radiography seems to be used to confirm ultrasound findings. Laboratory tests were used only occasionally. Generally they are not considered useful in diagnosing sinusitis (14), although in the Norwegian study GPs ranked c-reactive protein as the best aid in diagnosing sinusitis after history and clinical examination (1 1). In some Scandinavian countries diagnostic irrigations are considered possible and useful (2,3,15). Elsewhere in Europe and the United States irrigation is not a primary care procedure. In the present study irrigations were uncommon but apparently used for diagnostic purposes since Scad J Prim Health Care 1996; 14 they were performed without a preceding imaging examination, even when this would have been available. More than half of the irrigations were negative implying their diagnostic nature. The lack of clinical imaging did not increase diagnostic irrigations or control visits due to diagnostic uncertainty. X-ray and ultrasound facilities did not increase irrigations either. Irrigations accumulated only in the group for which both imaging facilities existed, and these patients were mostly treated during office hours when time-consuming procedures are easier to perfom. Our study highlights the difficulties of diagnosing maxillary sinusitis in primary care. Maxillary sinusitis was diagnosed in over 3/4 of all patients in whom the diagnosis was considered. When technology was not used, over 90% of patients were treated for sinusitis. The use of ultrasound findings in a similar way as a laboratory test with several possible cut-off points (16) illustrates also the willingness to err on the side of diagnosis. This suggests that the diagnostic criteria for sinusitis are very relaxed: suspicion of the disease almost justifies diagnosis and treatment. It is notable that, when diagnostic technology was not utilized, a previous sinusitis, which has a negative correlation to present sinusitis according to Axelsson (l), seemed to be considered an important factor in assessing the present illness and deciding whether diagnostic technology was needed. The variance in the frequency of sinusitis diagnoses between different diagnostic facility groups did not vary strongly. A tendency to rule out sinusitis was present when both x-ray and ultrasound could be used. Two Swedish studies (9,lO) from small health centres have shown similar trends. Molstad (10) confirmed a diagnosis of sinusitis in 73% of the patients who were examined with ultrasound, and the proportion of sinusitis among upper respiratory infections went down from 43% to 33%. Bjerre and Block (9) compared the number of sinusitis diagnoses in their health centre (with ultrasound) with three others with no such facility and detected only a tendency to a slightly lower incidence of sinusitis ( ). The availability of diagnostic technology increased general practitioners confidence in the correctness of their diagnoses. Diagnoses were uncertain in 61% of the cases when technology

7 did not exist. This diminished to 34% when both ultrasound and x-ray were available. In the Netherlands only 35% of the clinical diagnoses of sinusitis were considered uncertain by the general practitioners (8). The Dutch physicians greater confidence in their clinical diagnosis may be related to different definitions of sinusitis in these two countries (17). Conclusions In Finnish primary care, ultrasound was used much more widely than radiography in the diagnostic process. X-rays were mostly taken when symptoms persisted, or when other diagnostic procedures were inconclusive. The use of ultrasound diminished the numbers of sinusitis diagnoses. However, almost a third of the ultrasound examinations were difficult to interpret. Physicians seemed to regard the ultrasound resu!ts on an ordinal scale, instead of a binomial healthy-diseased scale. The techniques of using ultrasound and understanding the findings need improvement. The diagnostic criteria for sinusitis used in secondary care are not appropriate in primary health care, since both the diagnostic facilities and the severity of cases differ from the hospital environment. A detailed set of criteria for sinusitis in primary care, perhaps differentiating between serous and purulent cases, would facilitate better comparability of the results from different studies. The problem in primary health care are the patients in the early phase of sinusitis with mild symptoms and signs. Studies following the natural process of suspected sinusitis in primary care are needed to provide information about the predictive value of symptoms, signs, and the findings available through diagnostic imaging. References Axelsson A, Runze U. Symptoms and signs of acute maxillary sinusitis. ORL J Otorhinolaryngol Relat Spec 1976;38: Savolainen S, Ylikoski J, Jousimies-Somer H. Differential diagnosis of purulent and non-purulent acute maxillary sinusitis in young adults. Rhinol- Ogy 1989;27: Enquist S, Lundberg C. Akut sinusit - nl, hur och av vem bor den behandlas? (Acute sinusitis - when, how and by whom should it be treated? In Swedish.) L&artidningen 1986;83: Diagnoses sinusitis in primary care Berg 0, Carenfelt C. Analysis of symptoms and clinical signs in the maxillary sinus empyema. Acta Otolaryngol (Stockholm) 1988; 105: Jensen C, von Sydow C. Radiography and ultrasonography in paranasal sinusitis. Acta Radio1 1987;28: Kurien M, Raman R, Job A. Roentgen examination of maxillary sinus, antral puncture and irrigation - a comparative study. Singapore Med J 1989;30: Revonta M. Ultrasound in the diagnosis of maxillary and frontal sinusitis. Acta Otolaryngoi Suppl 1980;370: Van Duijn NP, Brouwer HJ, Lamberts H. Use of symptoms and signs to diagnose maxillary sinusitis in general practice: comparison with ultrasonography. BMJ 1992;305: Bjerre B, Block P. Diagnostik och behandling av maxillarsinusiter i primw4rd. ( and treatment of maxillary sinusitis in primary health care. In Swedish.) Allmanmedicin 1989; 10: Molstad C. Ultraljudundersokning av sinus - ett exempel fran en liten vhdcentral. (Ultrasound examination of sinus, an example from a small health centre. In Swedish.) Allmiinmedicin 1991;12: Lindbaek M, Hjortdahl P. Sinusitt i allmennpraksis - en diagnostisk utfordring. (Sinusitis in general practice - a diagnostic challenge. In Norwegian.) Tidsskr Nor Laegeforen 1993; 113: Boerma WGW, de Jong FAJM, Mulder PH. Health care and general practice across Europe. Utrecht: NIVEL, The Netherlands Institute of Primary Health Care, Mikela M, Sintonen H. Rationality and cost-effectiveness of diagnosis and treatment of group A streptococci in primary care patients with pharyngitis. Scand J Infect Dis 1991;23: Weymuller EA, Rice DH. Surgical management of infectious and inflammatory disease. In: Cummings GW, Fredrikson JM, Harker LA, Krause CJ, Schuller DE. Otolaryngology. Head and neck surgery. St.Louis: Mosby, 1993: Virtanen H. Sinuiitin hoito. (Treatment of sinusitis. In Finnish.) Duodecim 1989;105:192@ Sackett DL, Haynes RB, Guyatt GH, Tugwell P. The interpretation of diagnostic data. In: Clinical epidemiology. A basic science for clinical medicine. Second Edition. Boston: Little, Brown & Co, 1991 ; De Bock G. NHG Standaard Sinusitis (NHG Standards of Sinusitis. In Dutch.) Huisarts Wet 1993;8: Received December 1994 Accepted June 1995 Scand 3 Prim Healrh Care 1996; 14